Comparative analysis of classification algorithms performance for statistical based intrusion detection system
Muhammad Junaid Muzammil, Sameer Qazi, Taha Ali · 2013
Identification of applications running on a high speed network is a critical task for the protection of the local network from intrusive and malicious traffic. Network Intrusion Detection Systems (NIDS) are used for the detection of intrusive and malicious traffic in computer networks. Signature based IDS have been the most accurate among the available NIDS. But this accuracy is achieved at a cost of intensive computational power and inability to detect exploits whose signatures do not exist in the database. Statistical based IDS present a way to overcome these drawbacks with a compromise over accuracy. The key to the accuracy of the Statistical IDS lies in the classification algorithm adopted and the training data set used. This work presents a comparison of different classifier adopted for Statistical IDS. The performance of these classifiers has been measured over Weka software against variable amount of training data set.